Hewlett Packard Enterprise (HPE) with the German Centre for neurodegenerative diseases (DZNE) works, that will use its new computer architecture “The machine” in brain research. Computer in Palo Alto as well as development systems at the DZNE should speed up the research and provide more precise results through the analysis of extremely large amounts of data.
in the Research Centre enormous amounts of data incurred in taking pictures of the brain by MRI (magnetic resonance imaging), whose processing makes troublesome limitations with traditional computer infrastructure. Elaborate measurements up to 0.5 GB of raw data can in the second — and thus incurred up to 2 TByte per hour – which are not completely save, but are discarded after conversion to images with far less information content. Remedy to a very large memory sizes bring here HPEs new computer architecture with “Memory-Driven Computing” – as its core, and the processor is no longer used.
“the machine”, the HPE CeBIT shows the first prototype on the as the common past 60 years computer architectures to replace memory centric architecture. This radically new architecture, specialized processors over optical connections with a universal memory communicate on certain workloads. The separation of main memory and mass storage is therefore being repealed. HPE uses Memristoren for the new store. As a result of simulations the company assumes higher processing speed for many workloads when compared to today’s machines a 8000 times.
in cooperation with the DZNE should be with pilot applications show how “The Machine” is to use the faster and more comprehensive calculation of results. As further examples, where many TB fall data, Hewlett Packard enterprise called the creation of genetic data and the automated microscopy of living cells.
scientists at the DZNE hope for totally new insights into the causes of Alzheimer’s disease and other dementias. The new architecture could prove as an ideal platform, the MRI data completely to store and to process – more descriptive images with potentially large benefits of the research, for the whole diagnostic practice.